On 2020-03-30 20:10, Andres Freund wrote:
Also, shouldn't dynahash be adjusted as well? There's e.g. the
following HASH_ENTER path:
/* report a generic message */
if (hashp->isshared)
ereport(E
On 2020-04-03 19:55, Robert Haas wrote:> I think this patch needs to be
reverted. The only places where it
changes anything are places where we were about to throw some error
anyway. But as Andres's analysis shows, that's not nearly good enough.
OK, reverted.
On Mon, Mar 30, 2020 at 2:10 PM Andres Freund wrote:
> One important issue seems to me to be the size of the array that
> TransactionIdIsInProgress() allocates:
> /*
> * If first time through, get workspace to remember main XIDs in. We
> * malloc it permanently to avoid r
On 2020/03/31 3:10, Andres Freund wrote:
Hi,
On 2020-03-30 19:41:43 +0900, Fujii Masao wrote:
On 2020/03/30 16:58, Peter Eisentraut wrote:
Improve handling of parameter differences in physical replication
When certain parameters are changed on a physical replication primary,
this is commun
Hi,
On 2020-03-30 19:41:43 +0900, Fujii Masao wrote:
> On 2020/03/30 16:58, Peter Eisentraut wrote:
> > Improve handling of parameter differences in physical replication
> >
> > When certain parameters are changed on a physical replication primary,
> > this is communicated to standbys using the X
